Taking the Open Service Definition to 1.0

[Iain Emsley-2]

Dear all,

Comments on Rufus's specifics below but I have a question. Should there
be a definition about the methodologies of any data / services offered?

Just thinking of any service that might pull in data from a social
network (the "friends" may not know about their data being pulled in to
another service and may not agree - an ongoing argument I believe) or
using social science or, for instance, medical or other scientific data,
amongst other forms. If a SaaS is offering a service based on data, it
must be in the user's interests in knowing the source of any data. If
you cannot verify the data or the source, any results based thereon must
be suspect because an end user cannot follow the bread crumbs. Similarly
if data is manipulated in any fashion, the methodolog(y/ies) must be
announced to the end user and, again, the source be present to trace the
original data to prove or re-run any tests or to test the
methodolog(y/ies).
> 1. Naming: should it be Free/Open Service Definition or just plain Open Service Definition?
>  
I prefer Open  Service Definition: it would at least cover any service
which may buy in data and free it up. I wonder whether the semantics of
free mightconfuse any users of either variety (i.e., the license or the
service).
> 2. URL: should we go for the short 'osd' or the longer
> 'open_service_definition' (or even just 'open_service' or
> 'free_open_service').
> Open Service Defintion might play more friendly with search engines and the fewer acronyms the better IMHO.
>
>  
Iain

Iain Emsley

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Mukundan R-2]

Dear all,

Support for simple "Open Service"

agree with these very valid points. Addressing these issues will truly
make it "open" which is better than "free" and why to confuse with a
"free/open"?

A data which has been "approved" by the provider for use in other places
and other ways is more open.

A few clarifications:
How do we define personal data?

The definition assumes API's to be by default Open. what would happen if
the source was LGPL based? Do API's still remain open?

thanks and regards
mukundan


On Wed, July 2, 2008 01:10, Iain Emsley wrote:

> Dear all,
>
> Comments on Rufus's specifics below but I have a question. Should there
> be a definition about the methodologies of any data / services offered?
>
> Just thinking of any service that might pull in data from a social
> network (the "friends" may not know about their data being pulled in to
> another service and may not agree - an ongoing argument I believe) or
> using social science or, for instance, medical or other scientific data,
> amongst other forms. If a SaaS is offering a service based on data, it
> must be in the user's interests in knowing the source of any data. If
> you cannot verify the data or the source, any results based thereon must
> be suspect because an end user cannot follow the bread crumbs. Similarly
> if data is manipulated in any fashion, the methodolog(y/ies) must be
> announced to the end user and, again, the source be present to trace the
> original data to prove or re-run any tests or to test the
> methodolog(y/ies).
>> 1. Naming: should it be Free/Open Service Definition or just plain Open
>> Service Definition?
>>
> I prefer Open  Service Definition: it would at least cover any service
> which may buy in data and free it up. I wonder whether the semantics of
> free mightconfuse any users of either variety (i.e., the license or the
> service).
>> 2. URL: should we go for the short 'osd' or the longer
>> 'open_service_definition' (or even just 'open_service' or
>> 'free_open_service').
>> Open Service Defintion might play more friendly with search engines and
>> the fewer acronyms the better IMHO.
>>
>>
> Iain
>
> Iain Emsley
>
> _______________________________________________
> okfn-discuss mailing list
> okfn-discuss@...
> http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss
>

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Rufus Pollock]

On 09/07/08 20:04, Mukundan R wrote:
> Dear all,
>
> Support for simple "Open Service"

Just when it looked like we had consensus! That said, I think that Open
Network Service Definition is still the front-runner. In the interests
of closing this out if I don't hear any more comments by tomorrow
evening I'm going to assume that overall everyone is happy to go with ONSD.

> agree with these very valid points. Addressing these issues will truly
> make it "open" which is better than "free" and why to confuse with a
> "free/open"?
>
> A data which has been "approved" by the provider for use in other places
> and other ways is more open.
>
> A few clarifications:
> How do we define personal data?

A good question. I guess we could go with how it is traditionally
defined as data which provides information about you and which you would
expect to not be provided to a third party without your permission.
However I'm sure we could do better -- though I think for the purposes
of the definition as it stands, given that personal data must be
provided to its 'owner', simple 'personal data' will suffice.

> The definition assumes API's to be by default Open. what would happen if
> the source was LGPL based? Do API's still remain open?

I'm not sure how LGPL would make a difference here. No one could make
the APIs proprietary without violating the underlying F/OSS licence.

Regards,

Rufus

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Patrick Anderson]

Sorry to chime in so late, but I have a question/concern about using
the word 'Network'.

What about a situation where the user is utilizing the code or data
stored on a machine that they do not own, but where they are sitting
at that same terminal?

For instance, I've thought of a "Free as in Freedom" cafe where
customers could rent time on computers to use Free Software or rent a
mini-theatre room where they could watch movies that are under a
license permitting such performances.

I would want those customers to be guaranteed an opportunity to make
"at cost" copies of that code and data, but notice it wouldn't be over
a network.

Thanks,
Patrick

On Thu, Jul 10, 2008 at 12:31 PM, Rufus Pollock <rufus.pollock@...> wrote:

> On 09/07/08 20:04, Mukundan R wrote:
>> Dear all,
>>
>> Support for simple "Open Service"
>
> Just when it looked like we had consensus! That said, I think that Open
> Network Service Definition is still the front-runner. In the interests
> of closing this out if I don't hear any more comments by tomorrow
> evening I'm going to assume that overall everyone is happy to go with ONSD.
>
>> agree with these very valid points. Addressing these issues will truly
>> make it "open" which is better than "free" and why to confuse with a
>> "free/open"?
>>
>> A data which has been "approved" by the provider for use in other places
>> and other ways is more open.
>>
>> A few clarifications:
>> How do we define personal data?
>
> A good question. I guess we could go with how it is traditionally
> defined as data which provides information about you and which you would
> expect to not be provided to a third party without your permission.
> However I'm sure we could do better -- though I think for the purposes
> of the definition as it stands, given that personal data must be
> provided to its 'owner', simple 'personal data' will suffice.
>
>> The definition assumes API's to be by default Open. what would happen if
>> the source was LGPL based? Do API's still remain open?
>
> I'm not sure how LGPL would make a difference here. No one could make
> the APIs proprietary without violating the underlying F/OSS licence.
>
> Regards,
>
> Rufus
>
> _______________________________________________
> okfn-discuss mailing list
> okfn-discuss@...
> http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss
>

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Masayuki Hatta-2]

Hi,

>>>>> In <48765597.4080500@...>
>>>>> Rufus Pollock <rufus.pollock@...> wrote:

> Just when it looked like we had consensus! That said, I think that
> Open Network Service Definition is still the front-runner. In the
> interests of closing this out if I don't hear any more comments by
> tomorrow evening I'm going to assume that overall everyone is happy
> to go with ONSD.

I don't want to be a sticker, but I'd rather like to vote for OSSD
(Open Software Service Definition).

Rationale:

0) As Patrick already mentioned, we are talking about "Software (as a)
Service" in general.  Network is not necessarily involved, even if
it's currently the most popular style of usage.  Face-to-face services
can be claimed as "Software Service".

1) The expression "Network Service" is kinda confusing, since it
sounds like some kind of ISPs or network providers.  I believe those
are not really the primal target of this definition.  "Software
Service" might be equally confusing, vague one, but AFAIK it is
generally used to indicate some kind of service providers which use
software.  I think some portion of those are the intended users of the
definition.

2) As Evan already mentioned, Open Network Service Definition is kinda
confusing since it's not clear whether this is about the Definition of
Open Network Services or Service Definition for Open Network(well,
OSSD might be confusing anyway.  This is the weakest argument).

3) Software = Data + Code (For Thorsten ;-)

4) OSSD looks similar to OSD.  It might introduce confusion, for sure,
but I think it also provides a sense of continuity.  Another "S" will
indicate that OSSD is one jump ahead of OSD ;-)

Those are my 2 cents. It's all up to you now.

Best regards,
MH

--
Masayuki Hatta
Graduate School of Economics, The University of Tokyo

http://www.mhatta.org/

mhatta@... / mhatta@...
mhatta@... / mhatta@... / mhatta@...

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Benj. Mako Hill-2]

<quote who="Patrick Anderson" date="Thu, Jul 10, 2008 at 12:51:59PM -0600">

> Sorry to chime in so late, but I have a question/concern about using
> the word 'Network'.
>
> What about a situation where the user is utilizing the code or data
> stored on a machine that they do not own, but where they are sitting
> at that same terminal?
>
> For instance, I've thought of a "Free as in Freedom" cafe where
> customers could rent time on computers to use Free Software or rent a
> mini-theatre room where they could watch movies that are under a
> license permitting such performances.

There's a slippery slope argument here that I think you want to avoid.
If you ask to use my computer for 5 minutes, it seems potentially
onerous -- and unreasonable -- that we treat this the same as
distribution or online use.

The impact on user freedom is about the role, degree, and amount of a
user's experience that is framed by a particular technology. Most of
technologies I can think of that I might be very worried about are
either distributed and possessed or used over a computer network.

I can think of some others (like Internet cafes that you allude to) but
I can't see how distributing source code and data for the people most
affected by these systems would help either user or developer freedom.
Maybe I just need to near a more threshed out example.

Regards,
Mako

--
Benjamin Mako Hill
mako@...
http://mako.cc/

Creativity can be a social contribution, but only in so far
as society is free to use the results. --GNU Manifesto


_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[John Bywater-2]

Benj. Mako Hill wrote:

<quote who="Patrick Anderson" date="Thu, Jul 10, 2008 at 12:51:59PM -0600">
  

Sorry to chime in so late, but I have a question/concern about using
the word 'Network'.
    

Me too. :-) ...

What about a situation where the user is utilizing the code or data
stored on a machine that they do not own, but where they are sitting
at that same terminal?

For instance, I've thought of a "Free as in Freedom" cafe where
customers could rent time on computers to use Free Software or rent a
mini-theatre room where they could watch movies that are under a
license permitting such performances.
    

There's a slippery slope argument here that I think you want to avoid.
If you ask to use my computer for 5 minutes, it seems potentially
onerous -- and unreasonable -- that we treat this the same as
distribution or online use.

The impact on user freedom is about the role, degree, and amount of a
user's experience that is framed by a particular technology. Most of
technologies I can think of that I might be very worried about are
either distributed and possessed or used over a computer network.

I can think of some others (like Internet cafes that you allude to) but
I can't see how distributing source code and data for the people most
affected by these systems would help either user or developer freedom.
Maybe I just need to near a more threshed out example.
  

If the move to "software-as-a-service" has followed from the "network", which supports distributing software functionality as a remote service, then the shorter name "software service" probably always implies "network", even where the network has "gone away". :-)

But maybe the chief distinction within the name "software service" isn't whether there is a network or not, perhaps that is assumed nowadays, but rather that the service isn't a professional service? At any rate, the name "network service" appears not to carry the same distinction.

One example of where a software service might not be considered "on the network" is where a new software service is being tested before being put into production. We may wish to test for conformity with various things before "putting it on the network" which in turn may never happen. A spurious example perhaps. :-)

Yet overall, I do feel the name "software service" is much more common than "network service", and as such a much better thing to conjoin with "open" and "definition".

J.



_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Rufus Pollock]

On 10/07/08 19:51, Patrick Anderson wrote:

> Sorry to chime in so late, but I have a question/concern about using
> the word 'Network'.
>
> What about a situation where the user is utilizing the code or data
> stored on a machine that they do not own, but where they are sitting
> at that same terminal?
>
> For instance, I've thought of a "Free as in Freedom" cafe where
> customers could rent time on computers to use Free Software or rent a
> mini-theatre room where they could watch movies that are under a
> license permitting such performances.
>
> I would want those customers to be guaranteed an opportunity to make
> "at cost" copies of that code and data, but notice it wouldn't be over
> a network.

Mako has already done a great job of summarizing why one would want to
be careful about widening the sense of 'service' here so I won't say
anything other than that I agree with him here.

~rufus

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Patrick Anderson]

On Thu, Jul 10, 2008 at 2:21 PM, Benj. Mako Hill <mako@...> wrote:

> If you ask to use my computer for 5 minutes,
> it seems potentially onerous -- and unreasonable -- that we treat this the same as
> distribution or online use.

This makes some sense, but I'm having trouble understanding where the
lines should be drawn.

What if I used your computer for 5 minutes *over a network*?

Let's say you have OpenOffice installed on your machine, and I just
want to use it to do a quick edit on a .doc file.

If I use some "remote control" software (even just X) to connect and
do that work, what should we require for that session to be considered
'open'?

What if the session lasts 1 hour?

What if I connect every day for 30 days?

When does the situation turn from 'personal' to 'public'?


> I can't see how distributing source code and data for the people most
> affected by these systems would help either user or developer freedom.

Well, let's say you are playing a video game at "Cafe Open".

You notice some things that bug you, and want to try your hand at fixing them.

If you are allowed "at cost" access to the sources (both code and
data), you might improve it to the benefit of the whole community.  If
not, then that work will not be performed.

Of course, you might be a terrible artisan, or be trying to cause
trouble, so those edits would need to be approved by the collective
owners of the hardware hosting that game, but I think you should have
"at cost" access the sources required for that production to at least
TRY.

Because of the value in allowing source access, I would also like to
make other machines available for "at cost" rental that are already
setup with development with tools like text editors, compilers,
debuggers for source, and 3D modeling, audio editors, image editors,
etc. for data.  So that any patron that wanted to help out would not
be closed out.

Sincerely,
Patrick

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Rufus Pollock]

The two options for naming remain:

   1. Open Network Service Definition
   2. Open Software Service Definition

Following recent posts the front-runner would now appear to be:

   Open Software Service Definition (OSSD)

In particular, the arguments regarding simplicity and greater
applicability made for the second option by John and others seem quite
compelling.

So if you've got a strong opinion one way or the other now is the time
to make your views known!

~rufus

On 11/07/08 12:28, John Bywater wrote:
> Benj. Mako Hill wrote:
>> <quote who="Patrick Anderson" date="Thu, Jul 10, 2008 at 12:51:59PM
>> -0600">
>>  
>>> Sorry to chime in so late, but I have a question/concern about using
>>> the word 'Network'.
>>>    
>
> Me too. :-) ...

[snip]


_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[evanprodromou]

Rufus Pollock wrote:

> The two options for naming remain:
>
>    1. Open Network Service Definition
>    2. Open Software Service Definition
>
> Following recent posts the front-runner would now appear to be:
>
>    Open Software Service Definition (OSSD)
>
> In particular, the arguments regarding simplicity and greater
> applicability made for the second option by John and others seem quite
> compelling.
>
> So if you've got a strong opinion one way or the other now is the time
> to make your views known!
>  

I'm fine with this.

-Evan


_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Rufus Pollock]

Great. It seems there is consensus. I've updated the page (and renamed
it) and I think we can now officially launch the Open Software Service
Definition (OSSD):

   <http://www.opendefinition.org/ossd/>

~rufus

On 14/07/08 08:01, Evan Prodromou wrote:

> Rufus Pollock wrote:
>> The two options for naming remain:
>>
>>    1. Open Network Service Definition
>>    2. Open Software Service Definition
>>
>> Following recent posts the front-runner would now appear to be:
>>
>>    Open Software Service Definition (OSSD)
>>
>> In particular, the arguments regarding simplicity and greater
>> applicability made for the second option by John and others seem quite
>> compelling.
>>
>> So if you've got a strong opinion one way or the other now is the time
>> to make your views known!
>>  
> I'm fine with this.
>
> -Evan
>

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Mike Linksvayer-2]

On Mon, Jul 14, 2008 at 5:37 AM, Rufus Pollock <rufus.pollock@...> wrote:
> Great. It seems there is consensus. I've updated the page (and renamed
> it) and I think we can now officially launch the Open Software Service
> Definition (OSSD):
>
>   <http://www.opendefinition.org/ossd/>

Congratulations!

Well, I guess it is too late, but I don't get "software service".
John Bywater wrote a few days ago 'Yet overall, I do feel the name
"software service" is much more common than "network service", and as
such a much better thing to conjoin with "open" and "definition".'

Not if you check authoritative sources, like Google and Wikipedia. :)

This is the first I've heard of "software service" to mean SaaS, as
opposed to some kind of professional service around software or a
daemon running locally.  By contrast, "network service" is crystal
clear.  Maybe SaaS will naturally collapse to "software service" and
this will be moot.  Hopefully this definition will be wildly
successful and help make that happen.

Mike

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Rufus Pollock]

On 14/07/08 16:56, Mike Linksvayer wrote:
> On Mon, Jul 14, 2008 at 5:37 AM, Rufus Pollock <rufus.pollock@...> wrote:
>> Great. It seems there is consensus. I've updated the page (and renamed
>> it) and I think we can now officially launch the Open Software Service
>> Definition (OSSD):
>>
>>   <http://www.opendefinition.org/ossd/>
>
> Congratulations!

Thanks -- though that is a general congratulations to the whole
community that has helped develop the ideas embodied in the definition.

> Well, I guess it is too late, but I don't get "software service".
> John Bywater wrote a few days ago 'Yet overall, I do feel the name
> "software service" is much more common than "network service", and as
> such a much better thing to conjoin with "open" and "definition".'
>
> Not if you check authoritative sources, like Google and Wikipedia. :)

You should have spoken up earlier -- though the danger would be we would
have never reached agreement :) Perhaps this is something we can keep
open for the v1.1 (or v2.0) which will undoubtedly be necessary as more
precise use-cases (and edge-cases) come in over time.

It does seem that none of the suggested names were perfect (for example
several people I've mentioned Open Network Services too think their
about rules for ISPs).

> This is the first I've heard of "software service" to mean SaaS, as
> opposed to some kind of professional service around software or a
> daemon running locally.  By contrast, "network service" is crystal
> clear.  Maybe SaaS will naturally collapse to "software service" and
> this will be moot.  Hopefully this definition will be wildly
> successful and help make that happen.

The key point is that we get the idea of 'freedom/openness' in relation
to services clearly out there together with the reasons why it is
important. As you say, while names are important they are certainly not
the be all and end all the crucial point is to get across the ideas they
represent (and to that end any label, if clearly understood, will be
adequate).

Regards,

Rufus

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[Mike Linksvayer-2]

On Tue, Jul 15, 2008 at 4:16 AM, Rufus Pollock <rufus.pollock@...> wrote:
> On 14/07/08 16:56, Mike Linksvayer wrote:
>> Not if you check authoritative sources, like Google and Wikipedia. :)
>
> You should have spoken up earlier -- though the danger would be we would
> have never reached agreement :) Perhaps this is something we can keep open
> for the v1.1 (or v2.0) which will undoubtedly be necessary as more precise
> use-cases (and edge-cases) come in over time.

Probably good I didn't -- "software service" is kind of growing on me.

> It does seem that none of the suggested names were perfect (for example
> several people I've mentioned Open Network Services too think their about
> rules for ISPs).
>
>> This is the first I've heard of "software service" to mean SaaS, as
>> opposed to some kind of professional service around software or a
>> daemon running locally.  By contrast, "network service" is crystal
>> clear.  Maybe SaaS will naturally collapse to "software service" and
>> this will be moot.  Hopefully this definition will be wildly
>> successful and help make that happen.
>
> The key point is that we get the idea of 'freedom/openness' in relation to
> services clearly out there together with the reasons why it is important. As
> you say, while names are important they are certainly not the be all and end
> all the crucial point is to get across the ideas they represent (and to that
> end any label, if clearly understood, will be adequate).

Yep.

Mike

_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss

[John Bywater-2]

Rufus Pollock wrote:

On 14/07/08 16:56, Mike Linksvayer wrote:
  

Well, I guess it is too late, but I don't get "software service".
John Bywater wrote a few days ago 'Yet overall, I do feel the name
"software service" is much more common than "network service", and as
such a much better thing to conjoin with "open" and "definition".'

Not if you check authoritative sources, like Google and Wikipedia. :)
    

You should have spoken up earlier -- though the danger would be we would 
have never reached agreement :)

Been away for a few days, but just to josh along with this a bit further, I wonder whether the concern Mike raises is a good one....

I assume: what is common is what is good; we're looking for a good name. :-)

So it's true, neither Wikipedia nor Google have a definition of "software service", which indicates fairly authoritatively the term isn't an official name. But that says nothing of how common it is. :-)

Without thinking about it much, the count of Google search results seems to be a measure of the circulation of a term, in other words how common it is. I didn't look before, but it turns out the word phrase "software service" has 32% more results on Google than "network service" (5.3m and 4m respectively) [1] [2].

[1] http://www.google.com/search?q=%22software+service%22
[2] http://www.google.com/search?q=%22network+service%22

At the same time, as Rufus has heard (mentioned by him below), Wikipedia definitively indicates that a network service is basically something your ISP provides [3] in other words bandwidth and network access, and not really anything to do with the application layer at all.

[3] http://en.wikipedia.org/wiki/Network_service_provider

For me, the best argument for OSD => OSSD would follow from the authoritative observation of Grady Booch that the history of software is increasing layers of abstraction. In this case, it is the service level which abstracts pure software functionality from the software system and its ever-attendant development/deployment/maintenance/migration process. But that has raised new concerns about access and ownership. Such concerns are addressed by this defintion, which is it's entire purpose. Injecting 'S-for-service' into the F/O-S-D name stack supports such a meaning, as Masayuki Hatta articulated well under the notion of continuity in his last post to this list.

 Perhaps this is something we can keep 
open for the v1.1 (or v2.0) which will undoubtedly be necessary as more 
precise use-cases (and edge-cases) come in over time.
  

Perhaps it would be useful to clarify in the definition somewhere that the name "Open Software Service Definition" can be read as:

"Open Software [as a] Service Definition"

OR

"Open Software [Application] Service Definition" ?


Incidentally, was there ever an "Open Application Service Definition" option in this discussion? The term Application goes back to the standard reference model. But still, despite being u:ber-official, it might not be so common anymore....

It does seem that none of the suggested names were perfect (for example 
several people I've mentioned Open Network Services too think their 
about rules for ISPs).
  

Indeed.

J.



_______________________________________________
okfn-discuss mailing list
okfn-discuss@...
http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss